Proc. Natl. Acad. Sci. USA Vol. 95, pp. 9047–9052, July 1998 Anthropology, Evolution Testing migration patterns and estimating founding population size in Polynesia by using human mtDNA sequences (hypervariable region Iyhuman evolutionyNew Zealand Maoriysensitivity analysis) ROSALIND P. MURRAY-MCINTOSH*†,BRIAN J. SCRIMSHAW‡,PETER J. HATFIELD§, AND DAVID PENNY* *Institute for Molecular BioSciences, Massey University, P.O. Box 11222, Palmerston North, New Zealand; ‡Institute of Environmental Science and Research, Wellington, New Zealand; and §Renal Unit, Wellington Hospital, Wellington, New Zealand Communicated by Patrick V. Kirch, University of California at Berkeley, Stanford, CA, March 30, 1998 (received for review July 10, 1997) ABSTRACT The hypervariable 1 region of human mtDNA Pre-Polynesians are thought to have occupied the eastern islands shows markedly reduced variability in Polynesians, and this of South-East Asia at '2,000 BC, their Lapita culture with its variability decreases from western to eastern Polynesia. Fifty- characteristic pottery expanding rapidly through Melanesia and four sequences from New Zealand Maori show that the mito- out from the Solomon Islands into the western islands of Polyne- chondrial variability with just four haplotypes is the lowest of any sia (such as Tonga and Samoa) by '1,200–1,000 BC (19, 20). sizeable human group studied and that the frequency of haplo- Expansion to the eastern islands of the Pacific occurred largely types is markedly skewed. The Maori sequences, combined with between AD 1 and 1,000 (20, 21). Eastern Polynesia includes 268 published sequences from the Pacific, are consistent with a Tahiti (Society Islands), Easter Island, Hawaii, Marquesas, series of founder effects from small populations settling new Northern and Southern Cooks, the Australs, Pitcairn, and New island groups. The distributions of haplotypes were used to Zealand. A colonizer model of rapid exploration and settlement estimate the number of females in founding population of New of the uninhabited eastern zone has been proposed based on Zealand Maori. The three-step simulation used a randomly searching by sailing upwind with a relatively safe downwind selected founding population from eastern Polynesia, an expan- return (19). sionary phase in New Zealand, and finally the random selection Polynesian populations are relatively homogeneous phenotyp- of 54 haplotypes. The results are consistent with a founding ically and genetically. Because they moved into unoccupied areas population that includes '70 women (between 50 and 100), and of the Pacific, for some 3,200 years they were less affected by sensitivity analysis shows that this conclusion is robust to small admixture with other populations. Strong effects from the small changes in haplotype frequencies. This size is too large for size of founder populations and genetic drift are expected, as well models postulating a very small founding population of ‘‘cast- aways,’’ but it is consistent with a general understanding of as environmental and cultural selection (22). Population bottle- Maori oral history as well as the results of recent canoe voyages necks have been inferred from minisatellite data (23). Within recreating early trans-oceanic voyages. eastern Polynesia, the last major settlement was New Zealand (Aotearoa) with archaeological sites dating back only to 800 BP (24) although there may be, as judged by release of the Pacific The hypervariable region 1 of human mtDNA now is used food-rat (Rattus exulans), earlier human contact by 1,800 BP (25). extensively to study the origins and migration of modern humans, Hawaii, Easter Island, and Aotearoa (New Zealand) are at the Homo sapiens sapiens (see, for example, refs. 1–3). Relatively end of chains of migrations. For the settling of Aotearoa, there consistent patterns of human origin and dispersal are emerging, but it is desirable to investigate in more detail the processes that are two major theories that represent ends of a continuum. The have led to the current global distribution. most widely accepted theory is a planned settlement with multiple Polynesia is the most suitable place to study human migration voyages over time bringing people, animals, plants, and cultural into previously uninhabited regions. The earliest migrations are artefacts. This fits with our understanding of Maori oral traditions recent (3,200–800 BP) (4, 5), and therefore evidence of migration (elaborated by 19th century ethnologists) describing epic voyages patterns and their consequences are easier to find. For example, requiring great navigational skills. Long sea voyages of Polyne- most of the earliest archaeological sites were not submerged by sians to New Zealand are supported by recent canoe voyages sea level rises at the end of the last ice age and common patterns using only traditional navigational knowledge; these voyages have of extinctions can be found (6). Because the arrival of Polynesians supported the oral tradition on such matters as seasonal timing, is so recent, it is relatively easy to sequence DNA from bones of the setting of courses, and pohutukawa trees (Metrosideros)in animals that have become extinct since human impact (7, 8), and bloom at arrival (16, 26). Traditional stories of return voyages DNA can be sequenced from old human bones in the area, such from New Zealand to eastern Polynesia are supported by the as on Easter Island (9). Polynesians are also well known from both discovery of New Zealand obsidian in the Kermadec Islands (27). anthropological and archaeological viewpoints (see, for example, Within this framework, McGlone et al. (28) propose that up to refs. 5 and 10–12) as well as genetically (see refs. 13 and 14). In 500 settlers arriving over several generations around 800 BP addition, the migrations are retained in the strong oral histories would have provided both the necessary numbers to occupy the of Polynesian people, and it has been possible to replicate, based early coastal sites found throughout the country and the neces- on traditional knowledge and skills, long canoe voyages between sary skills, traditions, and knowledge for successful colonization. major island groups (15, 16). This number provides the basis for an initial rapid population Polynesians are speakers of a subgroup of Austronesian lan- growth expected in an environment previously uninhabited and guages, and their origins are suggested to extend back into therefore extremely rich in seals, large birds (including flightless mainland Asia (17, 18) with some Melanesian genetic admixture. moa), fish, and shellfish. Population from '600 BP is postulated to have grown more slowly to an estimate of 115,000 at 300 years The publication costs of this article were defrayed in part by page charge ago (12). payment. This article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: BP, before present. © 1998 by The National Academy of Sciences 0027-8424y98y959047-6$2.00y0 †To whom reprint requests should be addressed. e-mail: rosmm@ PNAS is available online at http:yywww.pnas.org. compuserve.com. 9047 Downloaded by guest on September 26, 2021 9048 Anthropology, Evolution: Murray-McIntosh et al. Proc. Natl. Acad. Sci. USA 95 (1998) At the other end of the spectrum are theories suggesting that chain termination technology (36) by using either the Sequenase islands were settled accidentally by random voyages made by one version 2.0 kit (United States Biochemicals), the Amplicycle canoe with perhaps 10–20 people (or at most a few canoes) sequencing kit (Perkin–Elmer), or an automated sequencing unit drifting before the wind or current–or by people forced to leave (Applied Biosystems model 373) of the Centre for Gene Tech- their home island, blindly searching for a new island on which to nology, University of Auckland. Because length heteroplasmy in settle (29). Other authors (see, for example, ref. 30) also restrict a G-C rich section in the control region of most Polynesian the number of original canoes by suggesting that the oral tradition mtDNA generates blurred sequence readout subsequently (37), of founding canoes really includes later subsidiary voyages within only the ‘‘L’’ strand of each sample was sequenced. Changes in New Zealand. Such theories are parsimonious with respect to the sequence at bases numbered lower than '16,189 therefore are number of long ocean voyages (see discussion in ref. 16) and have not reported. Each sample was repeated two or more times until had some acceptance as minimalist theories by which to interpret there were no ambiguities in base assignment. cultural development in isolation, without multiple interchange of A three-step simulation based on the observed frequencies of people, ideas, language, artefacts, animals, and plants. haplotypes (Table 1) in eastern Polynesia (2, 3) and in New It is now possible to test these models by using mtDNA. The Zealand (ref. 3; this study) was used to estimate the numbers of expectation in the settlement of Pacific Islands is a series of females founding the Maori population in New Zealand. The first founder effects, modified by immigration and emigration over step selected randomly, with replacement, a maternal founding several generations. Under this model, Maori and Hawaii are population of from 4 to 250 people from the frequencies of the expected to be a subset of the genetic variability in central eastern 11 haplotypes observed in eastern Polynesia (Table 1, column Polynesia,